(3) Number of Sample Replications . In choosing the number of 

 replicates to take at each station, the practical limitations of time 

 and resources, the nature of the research objectives, and the results of 

 the preliminary analysis should also be considered. In Oliver and 

 Slattery (1976), eight replicates contained 87 percent of the total number 

 of individuals. The confidence limits (Table 3) indicate the accuracy 

 of the abundance estimates with the various numbers of replicate samples. 

 Eight samples contained 76 percent of the total number of species in the 

 28 samples and additional replicates added new species at a much slower 

 rate (Fig. 2). Also, diversity, H", and evenness, J, values, commonly 

 used for population analysis (Sec. VII), appear to stabilize between four 

 and eight replicate samples (Fig. 3). Therefore, eight appears to be the 

 optimum number of replicates to estimate the number of species based on 

 the additional cost and manpower that would be required to increase the 

 replications (see Sec. VIII). 



12 14 16 

 Replicates 



2 4 6 8 10 12 14 16 !8 20 22 24 26 28 

 Replicates 



Figure 3. Species diversity, H", and evenness, J, values 



calculated from increasing numbers of replicate samples 

 (Oliver and Slattery, 1976). 



4. Distribution of Samples . 



The detailed patterns of distribution of the species sampled are 

 not examined in this report. However, to construct a quantitative 



